Our long-term objective is to ascertain how protein conformation plays a role in biological function and in various diseases. Our specific aims are to finish our development of our physics-based united-residue (UNRES) approach to the protein folding problem, i.e., to compute structure, folding pathways, and thermodynamic and dynamic properties. This involves replacing the last remaining knowledge-based term, corresponding to side chain-side chain interactions, by physics-based terms, extension of UNRES to simulate folding of disulfide-containing proteins, and to treat the lipid-membrane environment. At the all-atom level, we will treat the pH-dependent ionization of side chains (including solvation), and the use of 13C1 chemical shifts in protein-structure simulation. We will continue the development of our UNRES model of nucleic acids (NA-UNRES) and merge UNRES and NA-UNRES into a viable package, which will be provided to the community. We will also continue the developments of sampling techniques and parallelization of UNRES/MD to carry out simulations of very large single-chain and oligomeric proteins and their complexes, and develop tools, based on Principal Component Analysis (PCA) for the analysis of mesoscopic-dynamics trajectories. We will demonstrate how these aims can lead to valid predictions of structures and folding pathways of proteins, and protein-nucleic acid and protein-protein complexes. Our main focus will then involve the application of this methodology to a biological problem: the mechanism of action of the human HSP70 chaperone.